Controlling chronic Pseudomonas aeruginosa infections by strategically interfering with the sensory function of SagS

SUMMARY

The hybrid sensor SagS plays a central role in the formation of P. aeruginosa biofilms, by enabling the switch from the planktonic to the biofilm mode of growth, and by facilitating the transition of biofilm cells to a highly tolerant state. In this study, we examined the importance of the SagS key amino acid residues associated with biofilm formation (L154) and antibiotic tolerance (D105) in P. aeruginosa virulence. Recombinant P. aeruginosa ΔsagS and ΔsagS chromosomally expressing wild-type sagS, or its two variants D105A and L154A, were tested for their potential to form biofilms and cause virulence in plant and mouse models of acute and chronic pneumonia. Although mutation of sagS did not alter P. aeruginosa virulence during acute infections, a significant difference in pathogenicity of sagS mutants was observed during chronic infections, with the L154A variant showing reduced bacterial loads in the chronic pneumonia model, while interference with the D105 residue enhanced the susceptibility of P. aeruginosa biofilms during tobramycin treatment. Our findings suggest that interference with the biofilm or tolerance regulatory circuits of SagS affects P. aeruginosa pathogenicity in chronic but not acute infections, and reveal SagS to be a promising new target to treat P. aeruginosa biofilm infections.

In Pseudomonas aeruginosa, biofilm formation and antibiotic tolerance are regulated by the hybrid sensor SagS via two independent pathways. However, how both of these pathways contribute to virulence has not been demonstrated before. Although mutation of sagS did not alter virulence during acute infections, here we show for the first time that biofilm formation and biofilm drug tolerance play distinct roles in the chronic virulence phenotype of P. aeruginosa.

Abbreviated Summary

Biofilms are the root cause of chronic infections. While previous studies indicated the ability to form biofilms to be linked to the acute-chronic switch in virulence, here we show for the first time that biofilm formation and biofilm drug tolerance play distinct roles in the chronic virulence phenotype of P. aeruginosa.

INTRODUCTION

The opportunistic pathogen Pseudomonas aeruginosa is a Gram-negative γ-proteobacterium that can infect a variety of organisms including plants, animals, and most importantly, humans. In particular, immunocompromised patients, burn-wound patients, and individuals suffering from cystic fibrosis are at risk (Hauser et al., 2011, Kerr & Snelling, 2009, Lyczak et al., 2002, Rosenberg et al., 2001, Pirnay et al., 2009, Rabello et al., 2015). During initial colonization, P. aeruginosa utilizes acute virulence factors including the type III secretion system (T3SS) (Hauser, 2009), proteases (LasA and LasB) (Bainbridge & Fick, 1989, Thibodeau & Butterworth, 2013, Kuang et al., 2011, Kessler et al., 1993), hydrogen cyanide (HCN) (Blumer & Haas, 2000, Lenney & Gilchrist, 2011), type IV pili (Kearns, 2010), flagella (Doyle et al., 2004), rhamnolipids (Abdel-Mawgoud et al., 2010, Alhede et al., 2009), phenazines (Briard et al., 2015, Mavrodi et al., 2001, Wang et al., 2011), and Pseudomonas Quinolone Signal (PQS) (Diggle et al., 2007). Many of these virulence factors are dependent on quorum sensing (QS) (Jimenez et al., 2012). P. aeruginosa harbors three major QS systems: las, rhl, and pqs that are regulated in an hierarchical manner with LasR activating the rhl system, while pqsR being positively regulated by LasR and negatively by RhlR (Jimenez et al., 2012, Wade et al., 2005). As infection progresses, P. aeruginosa switches from a planktonic to a biofilm mode of growth, characterized by the formation of an extracellular matrix composed of exopolysaccharides, proteins and extracellular DNA (Costerton et al., 1995, Davey & O’Toole G, 2000, Mann & Wozniak, 2012). This matrix protects the bacteria from the immune system as well as antibiotics. Furthermore, biofilm cells are up to thousand times more resistant to antibiotics that their planktonic counterparts, which can be related to the tolerant state they obtain inside the biofilm (Luppens et al., 2002, Liao & Sauer, 2012, Liao et al., 2013). Therefore, it is not surprising that biofilm infections, which have been estimated to include 80% of all human bacterial infections, and 90% of chronic wound infections (Römling & Balsalobre, 2012, Attinger & Wolcott, 2012), are highly recalcitrant to traditional therapies. However, while factors contributing to acute infections are well known, little is known about how the biofilm mode of growth contributes to chronic infections.

The transition from the planktonic life style to the surface associated mode of growth and subsequent formation of biofilms is governed by a series of regulatory signaling cascades (Valentini & Filloux, 2016, Winstanley et al., 2016). In P. aeruginosa, the switch to the biofilm mode of growth involves the sensor-regulator hybrid SagS. Inactivation of sagS coincides with biofilm development being arrested at the irreversible attachment stage and attached cells remaining susceptible to antimicrobial agents (Petrova & Sauer, 2011, Petrova et al., 2017, Sauer et al., 2002, Petrova & Sauer, 2009). More specifically, SagS regulates (i) the transition to biofilm formation and (ii) the transition to cells gaining their enhanced tolerance to antimicrobials via two distinct mechanisms, with biofilm formation requiring phosphosignaling between SagS and the two-component system BfiSR (Petrova & Sauer, 2009, Petrova & Sauer, 2010, Petrova & Sauer, 2011, Petrova et al., 2017). In contrast, biofilm drug tolerance was found to be independent of phosphotransfer, biomass accumulation, and biofilm architecture, but dependent on BrlR (Gupta et al., 2013, Gupta et al., 2014, Petrova et al., 2017). BrlR is a c-di-GMP-responsive transcriptional regulator that acts downstream of SagS and activates the expression of an array of multidrug efflux pumps and ABC transporters (Liao & Sauer, 2012, Liao et al., 2013, Chambers et al., 2014, Poudyal & Sauer, 2018), with the ABC transporter PA1874–77 directly contributing to the antibiotic tolerance of biofilms (Poudyal & Sauer, 2018).

SagS is composed of three domains: an N-terminal periplasmic sensory domain (HmsP), a histidine kinase domain (HisKA) and a C-terminal phosphoreceiver (Rec) domain (Fig. 1). The periplasmic sensory domain is flanked by two transmembrane helices and transmits the signal to the HisKA and Rec domains via conformational changes in the HAMP domain (Fig. 1, (Dingemans et al., 2018b)). SagS contributing to two independent pathways is made possible by its domain composition. Biofilm formation is enabled in a manner dependent on the HisKA and Rec domains while transition to cells gaining their enhanced drug tolerance required the presence of the periplasmic HmsP sensory domain (Fig. 1, (Petrova et al., 2017)). Furthermore, recent findings suggest the presence of two distinct sets of amino acid residues that when mutated, block one of the two sensory functions of SagS. For example, substitution of amino acid residue L154 with alanine rendered P. aeruginosa unable to develop mature biofilms while alanine replacement of D105 coincided with biofilms that were susceptible to antimicrobial agents (Dingemans et al., 2018b). In addition to biofilm-associated phenotypes, SagS furthermore contributes to the virulence (Petrova & Sauer, 2011). This was supported by the finding of sagS inactivation rendering P. aeruginosa significantly more virulent than wild type using an Arabidopsis plant model of infection.

Figure 1. Model of SagS regulatory pathways leading to biofilm formation and antibiotic tolerance.

Based on our previous findings, the membrane bound sensor SagS contributes to biofilm formation in a manner dependent on BfiSR and to biofilm drug tolerance via BrlR (Petrova & Sauer, 2011, Petrova et al., 2017, Sauer et al., 2002, Petrova & Sauer, 2009, Gupta et al., 2013, Gupta et al., 2014). Upon sensing an as of yet unidentified cue, a signal is transduced from the periplasmic HmsP sensory domain of SagS to its kinase domain via the HAMP domain, followed by phosphotransfer between the histidine at position 315 in the HisKA domain and the aspartic acid in position 713 in the phosphoreceiver domain, resulting in the subsequent phosphorylation of BfiSR. Residue L154 is part of the regulatory circuit that leads to biofilm formation via a cascade that starts with the phosphorylation of BfiS by SagS. In contrast, the key residue D105 is involved in drug tolerance via the indirect regulation of BrlR transcript levels in the presence of elevated concentrations of c-di-GMP. HmsP, sensory domain. HAMP, HAMP domain. HisKA, histidine kinase domain. Rec, phosphoreceiver domain. Transmembrane spanning helices are indicated by three vertical lines.

Considering that SagS contributes to virulence while also regulating biofilm formation and drug tolerance, we asked whether the biofilm formation or biofilm drug tolerance pathways contribute to the virulence phenotype of P. aeruginosa. To do so, we focused on two amino acid residues in the HmsP sensory domain of SagS that are involved in biofilm formation (L154) or antibiotic tolerance (D105), respectively (Fig. 1). To our knowledge, the sole contribution of biofilm formation or drug tolerance alone to infection and the pathogenicity phenotype of P. aeruginosa has never been demonstrated, as both phenotypes tend to be linked.

In this study, we show that: (1) Inactivation of sagS had little to no effect on virulence gene expression or pathogenicity of P. aeruginosa in models of acute infection. (2) ∆sagS mutants that are impaired in biofilm formation in vitro are defective for persistent, chronic infections in the murine lung. (3) The ability to form in vivo biofilms in the murine lung positively correlates with the SagS amino acid residue L154. (4) Biofilm formation rather than differences in virulence gene expression contribute to persistent, chronic infections. (5) The SagS-dependent pathway leading to biofilm drug tolerance (via amino acid residue D105) is essential for P. aeruginosa in vivo biofilms to resist antibiotic therapy. Moreover, our findings suggest that (6) targeting SagS by interfering with SagS signaling is a promising new target to treat P. aeruginosa biofilm infections.

RESULTS

Mutation of the key residue L154 in the sensory domain of SagS impairs biofilm formation.

We previously demonstrated that in Pseudomonas aeruginosa, the sensor-regulator hybrid SagS regulates both, (i) the transition to biofilm formation and (ii) the transition to cells gaining their enhanced tolerance to antimicrobials (Petrova et al., 2017, Dingemans et al., 2018b, Gupta et al., 2013, Petrova & Sauer, 2011). Moreover, we showed that individual amino acids located in the sensory HmsP domain of SagS contribute to SagS regulating the transitions, with amino acid residue L154 being required for biofilm formation while D105 contributed to biofilm drug tolerance (Dingemans et al., 2018b).

To determine the role of these two divergent regulatory circuits in P. aeruginosa virulence, we first eliminated the need of plasmid-borne expression of sagS and sagS variants and antibiotics for plasmid maintenance by constructing unmarked ΔsagS mutant strains that chromosomally express the wild-type sagS gene or its two variants, sagS_L154A and sagS_D105A, under the control of the native sagS promoter. Based on previous findings (Dingemans et al., 2018b), we anticipated complementation of ΔsagS with wild-type sagS and sagS_D105A, but not sagS_L154A, to restore biofilm formation. We therefore made use of the difference in biofilm architecture to confirm that the resulting chromosomal constructs were comparable to strains overexpressing sagS and sagS variants. Analysis of the biofilm architecture by confocal microscopy demonstrated that chromosomal expression of wild-type sagS and sagS_D105A restored biofilm formation by ΔsagS to wild-type levels, while chromosomal expression of sagS_L154A failed to do so (Fig. 2A). In contrast, biofilms formed by ΔsagS::sagS_L154A were comparable in architecture to biofilms by ΔsagS harboring the empty CTX (Fig. 2A). COMSTAT analysis confirmed biofilms by ΔsagS, empty vector control (>16-fold), and ΔsagS::sagS_L154A (>6-fold) to be significantly reduced in biofilm biomass relative to PAO1, while the biomass of biofilms by ΔsagS::sagS and ΔsagS::sagS_D105A was comparable to PAO1 (Fig. 2B). Analysis of the average and maximum thickness (Fig. 2C, Fig. S1) furthermore confirmed the visual differences in the biofilm architecture amongst the strains (Fig. 2A). Our findings are in agreement with previous findings (Dingemans et al., 2018b) and indicate that chromosomal expression of sagS and sagS variants under the control of the native sagS promoter results in biofilms of comparable architecture as strains expressing sagS and sagS variants making use of an overexpression vector system. Previous findings indicated inactivation of sagS in a P. aeruginosa PA14 background to coincide with significantly increased virulence 6 and 9 days post infection using an A. thaliana infection model relative to the parental strain (Petrova & Sauer, 2011). To further confirm that the chromosomal constructs were comparable to strains devoid of the CTX vector, we infected 8-day-old A. thaliana plants with the different ΔsagS mutant strains (in a PAO1 background) and monitored plant death over 2 weeks. In agreement with the previous findings obtained using a PA14 ΔsagS mutant (Petrova & Sauer, 2011), a PAO1 ΔsagS mutant strain harboring the empty vector was hypervirulent compared to a ΔsagS strain that chromosomally expressed the wild-type sagS gene (ΔsagS::sagS) (Fig. 3). A similar trend was observed when 12-day-old A. thaliana plants were used (Fig. S2). The findings indicated not only that SagS contributes to virulence in a similar manner in both PAO1 and PA14 despite differences in the gene content but that ΔsagS mutant strains harboring the empty vector were as hypervirulent as strains lacking the CTX vector.

Figure 2. Biofilm formation is impaired in ΔsagS, ΔsagS harboring the empty vector, and the L154A variant.

(A) Representative confocal microscopy images of 3-day-old biofilms by wild type, ΔsagS, ΔsagS harboring the empty CTX vector or chromosomally expressing sagS, sagS_L154A and sagS_D105A, under the control of the native sagS promoter. Biofilms were grown in 5-fold-diluted LB medium and stained prior to the LIVE/DEAD BacLight viability stain and image acquisition. White bars represent 100 μm. Total biofilm biomass (B) and biofilm thickness (C) were determined via COMSTAT analysis. COMSTAT data obtained using strains overexpressing sagS, sagS_L154A and sagS_D105A from the plasmid pMJT-1 were used for comparison. All assays were performed at least in duplicate, with a minimum of 6 images being captured. Error bars indicate standard deviation. *, significantly different (P<0.05) from PAO1.

Figure 3. Inactivation of sagS renders P. aeruginosa hypervirulent in an Arabidopis thaliana infection model.

(A) Survival curves of 8-day-old A. thaliana plants in the absence of bacteria (media control, black) or post infection with ΔsagS::CTX (red) or ΔsagS::sagS (green). Inset, Representative images of A. thaliana plants that were recorded as “dead” (left, all leaves are discolored) or “alive” (right). (B) Box plot representation of plant death in 8-day-old A. thaliana 9 days post infection with indicated ΔsagS mutant strains. All experiments were performed in triplicate, with at least 8 plants per strain per experiment. The median is indicated with a horizontal line, while the mean is indicated with a “+”. **, significantly different (P<0.01) relative to ΔsagS::sagS.

Inactivation of sagS only affects biofilm marker genes, but not acute virulence genes.

To determine the contribution of the two divergent regulatory SagS circuits in P. aeruginosa virulence, we first asked if mutations of sagS also affect the expression of virulence genes under planktonic and biofilm conditions using qRT-PCR. Under planktonic conditions, only minor differences in the expression of genes previously reported to contribute to acute virulence (Hauser, 2009, Bainbridge & Fick, 1989, Thibodeau & Butterworth, 2013, Kuang et al., 2011, Kessler et al., 1993, Blumer & Haas, 2000, Lenney & Gilchrist, 2011, Kearns, 2010, Doyle et al., 2004, Abdel-Mawgoud et al., 2010, Alhede et al., 2009, Briard et al., 2015, Mavrodi et al., 2001, Wang et al., 2011, Diggle et al., 2007, Jimenez et al., 2012) were detected (≤2-fold difference in transcript abundance, Fig. 4A,) Moreover, with the exception of exoS, no significant difference in transcript abundance between ΔsagS and strains expressing sagS or sagS variants was noted (Fig. 4A). However, under biofilm conditions, several biofilm marker genes such as pelA, pslG, glpD, anr, and dnr (Bielecki et al., 2013, Dingemans et al., 2016, Dingemans et al., 2018a, Son et al., 2007, Crabbe et al., 2010, Crabbe et al., 2008, Goodman et al., 2004) as well as genes contributing to motility (fliC, pilY1) were >2-fold down-regulated in biofilms by ΔsagS strain relative to ΔsagS::sagS (Fig. 4B). Interestingly, biofilms formed by ΔsagS::sagS_D105A and ΔsagS::sagS_L154A demonstrated intermediate levels of gene expression for most of the genes tested suggesting partial restoration of the ΔsagS phenotype (Fig. 4B).

Figure 4. Gene expression of planktonic (A) and biofilm (B) cells.

For gene expression analysis by qRT-PCR, RNA obtained from (A) planktonic cells grown to the exponential growth phase, and (B) biofilms grown for 3 days were used. Gene expression was normalized to the expression of ΔsagS chromosomally expressing sagS (ΔsagS::sagS), with cysD and mreB being used as housekeeping genes under planktonic and biofilm growth conditions, respectively. T3SS, Type III Secretion System. QS, Quorum Sensing. All assays were performed at least in triplicate. Error bars indicate standard deviation. *, significantly different (P<0.05) from ΔsagS::PsagS-sagS.

Motility is reduced in sagS mutants.

Given the intermediate level of expression by ΔsagS::sagS_D105A and ΔsagS::sagS_L154A relative to ΔsagS and ΔsagS::sagS including genes contributing to motility, we next asked whether the difference in gene expression was sufficient to have an effect on swimming and swarming motility. ΔsagS mutants were defective in swarming, with chromosomal expression of sagS or sagS variants restoring swarming motility (Fig. 5A). While no difference in swarming motility between ΔsagS::sagS and ΔsagS expressing the sagS variants was noted following 24 h of incubation, continued incubation revealed significant differences in the swarming motility by ΔsagS::sagS and ΔsagS expressing the sagS variants, with the ΔsagS::sagS_D105A and _L154A strains showing intermediate levels of swarming motility relative to ΔsagS and ΔsagS::sagS (Fig. 5A). However, no difference in swarming motility by strains expressing the sagS variants was noted. Moreover, no significant difference in swimming motility was noted in the presence or absence of sagS or sagS variants 48h post inoculation (Fig. 5B). It is important to note that the observed differences in motility were not due to defects in growth under the conditions tested, as all tested strains demonstrated similar growth rates in swarming (Fig. S3A) and swimming medium (Fig. S3B).

Figure 5. Swarming and Swimming motility of ΔsagS mutant strains.

(A) Swarming motility was determined using 0.4 agar swarming medium while (B) swimming motility assays made use of 0.3% agar. Measurements were taken from the inoculation point to 24h and 48h post-incubation. Insets, representative images showing swarming and swimming motility of the indicated mutant strains. All assays were performed at least in triplicate. Error bars indicate standard deviation. *, significantly different (P<0.05) from ΔsagS::PsagS-sagS.

SagS does not contribute to early signs of infection and lethality.

While the analysis of virulence gene expression suggested SagS to contribute little to acute virulence, inactivation of sagS had an effect on motility. Given that numerous studies have shown that virulence and motility are often intimately linked by complex regulatory networks (Josenhans & Suerbaum, 2002), we therefore asked whether SagS contributes to virulence using a romaine lettuce leaf infection model. The romaine lettuce leaf model was chosen as this model is known to show early signs of infection and lethality. However, no significant difference in plant pathogenicity was detected between the four sagS mutant strains tested, even after monitoring the spread of P. aeruginosa infection up to 3 days post infection (Fig. 6A). The findings of SagS not contributing to early signs of infection or lethality is in agreement with SagS contributing little to none to the expression of virulence genes.

Figure 6. Contribution of SagS to P. aeruginosa virulence and pathogenicity using two acute models of infections.

A lettuce model of virulence (A) and murine acute pneumonia model (B-D) were used. (A) Inactivation of sagS has no effect on P. aeruginosa pathogenicity using a lettuce model of virulence. The midrib of romaine lettuce leaves post infection with indicated ΔsagS mutant strains was monitored for 3 days, with spread of infected zone being measured daily for up to 3 days. Experiments were repeated at least in triplicate. (B-D) To determine whether inactivation of sagS renders P. aeruginosa less virulent, seven-week old CD1 mice were intranasally inoculated with ~10⁷ CFU/ml of ΔsagS and PAO1 strains (1:1 ratio), and the bacterial burden was determined 24 h post infection. The bacterial burden is expressed as log₁₀ CFU/lung. ****, significantly different (P<0.0001) compared to the parental strain PAO1. A cohort of 4 mice, equal number of males and females, was used. (B) Competitive index assay using seven-week old CD1 mice post. Bacterial burden present in the lungs of seven-week old CD1 mice post (C) 24 h and (D) 36 h of infection. Mice were intranasally inoculated with ca. 10⁷ CFU/ml of each P. aeruginosa strain. Cohorts of 10 mice, equal number of males and females, were used for each time point. Error bars indicate standard deviation. ****, significantly different (P<0.0001) from ΔsagS::sagS.

sagS inactivation coincides with reduced fitness and reduced virulence during later stages of acute infection.

We furthermore examine the ΔsagS mutant strains to establish an acute infection using competitive mixed infection assays. Competitive mixed infection assays have been widely used to assess the fitness and virulence of individual mutant strains of P. aeruginosa (Lau et al., 2004, Yoon et al., 2006), Vibrio cholerae (Freter et al., 1981, Taylor et al., 1987) and Salmonella enterica serovar Typhimurium (Beuzon et al., 2000, Shea et al., 1999) versus their parental strains during in vivo infection. Competitive mixed infection assays were carried out using a murine model of acute pneumonia. Wild-type and ΔsagS mutant bacteria were used to infect adult CD-1 mice (in groups of five) intranasally with 1 × 10⁷ cells (1:1 ratio). Following 24 hr, infected lungs were recovered for bacterial load determinations. A competitive index of 1 indicates that the two strains are proliferating equally in vivo. However, the ΔsagS mutant was only 17% as (or 6-fold less) competitive as its parental strain PAO1 (Fig. 6B), suggesting sagS inactivation coincides with reduced fitness.

Given the noted reduced fitness of ΔsagS compared to PAO1 in the absence of SagS contributing to early signs of infection or lethality (Fig. 6B), we further determined whether the reduced fitness was due to differences in growth by ΔsagS, PAO1, ΔsagS harboring the empty vector, wild-type sagS as well as the D105A and L154A variants. No difference in growth rate was observed between the different strains when grown in LB medium (Fig. S3C). We furthermore compared the growth using M9 medium containing 0.05% phosphatidylcholine as the sole carbon source. This is based on the finding that P. aeruginosa can utilize phosphatidylcholine as a major nutrient source in the lung environment (Son et al., 2007, Sun et al., 2014). Similar to growth in LB, however, no difference in growth rate was observed between the different strains tested when grown in M9 medium containing 0.05% phosphatidylcholine (Fig. S3D).

The findings clearly suggested that the reduced fitness by ΔsagS relative to PAO1 was not due to differences in growth. To determine how the reduced fitness by ΔsagS affects the establishment of acute infections in the murine model of acute pneumonia, single infection studies were carried out. No difference in virulence was noted using a murine model of acute pneumonia 24h post infection (Fig. 6C). Continued infection, however, revealed differences in virulence. Following 36h post infection, the bacterial load of the isogenic ΔsagS strain harboring the empty vector was reduced by >10-fold as compared to infection with the sagS mutant strains expressing sagS or sagS variants (Fig. 6D). The findings suggested that SagS likely contributes to fitness in vivo only during the later stages of acute infection. Considering that only inactivation of sagS, but not alanine substitutions of L154 and D105 that contribute to biofilm formation and drug tolerance, respectively, affected the late stage of acute infections, our findings furthermore suggest that either biofilm formation or drug tolerance alone are sufficient during late stage acute infections.

Interference with residue L154 in the sensory domain of SagS leads to a significantly reduced bacterial burden by P. aeruginosa in a murine chronic pneumonia model.

Given the observation that a ΔsagS mutant shows reduced virulence after 36 h of infection in an acute pneumonia model and that SagS is required for biofilm formation, with biofilms having been linked to chronic infections, we next asked whether the divergent SagS regulatory pathways play a role in long-term infections. We therefore transitioned to a vertebrate model of chronic infection. Specifically, we made use of a chronic pneumonia model established by Cash et al. (Cash et al., 1979) except that we made use of a murine rather than a rat model. The model is based on the intratracheal administration of agarose beads impregnated with P. aeruginosa. This model was chosen because it allows the study of chronic infection, marked by the formation of persistent bacterial biofilm populations at the site of infection that can persist for up to 6 months and mimic biofilm-related infections (Lau et al., 2005, Lau et al., 2004, Yoon et al., 2006, Garcia-Medina et al., 2005, Woods et al., 2005). Considering that ΔsagS and ΔsagS::sagS_L154A were found to be impaired in biofilm formation in vitro, we expected the two mutant strains to have a disadvantage in establishing biofilms and/or persistent infection compared to the complemented ΔsagS::sagS strain. In agreement with in vitro findings, following 3 days of infection, the bacterial load in the lungs of mice infected with ΔsagS was significantly lower (2.5-log decrease) than that of complemented ΔsagS::sagS strain (Fig. 7A). Moreover, compared to the initial inoculum (1.2 × 10⁶ CFU), more than a 100-fold reduction was noted for ΔsagS while a slight increase in the bacterial burden relative to the inoculum was noted 3 days post infection for the complemented ΔsagS::sagS strain. Similar to ΔsagS, the bacterial load by ΔsagS::sagS_L154A in the lungs of mice infected was significantly reduced (2.5-log decrease) relative to the ΔsagS::sagS (Fig. 7A). In contrast, no significant difference in the bacterial load in the murine lungs infected with ΔsagS::sagS and ΔsagS::sagS_D105A was noted (Fig. 7A). The difference in the bacterial load among the ΔsagS mutant strains persisted at day 5 post infection (Fig. 7B). Considering that ΔsagS and ΔsagS::sagS_L154A, but not ΔsagS::sagS_D105A, were found to be impaired in biofilm formation in vitro, our findings suggest that bacterial load and thus, persistence by P. aeruginosa in the murine lung correlates with in vitro biofilm formation phenotypes. Our findings furthermore suggest persistence and thus, virulence, correlates with the ability to form biofilms but to be independent of drug tolerance.

Figure 7. The bacterial burden positively correlates with the SagS amino acid residue L154 while biofilm tolerance positively correlates with the residue D105 using a murine chronic pneumonia model.

Seven-week old BALB/c mice (cohorts of 10; equal number of males and females) were intratracheally inoculated with individual P. aeruginosa strains using a ball-ended needle. The bacterial burden was subsequently determined (A) 3 days and (B) 5 days post infection. In addition, 24 h post infections, separate cohorts of infected mice were treated once daily with 50 mg/kg tobramycin (TOBI), and the bacterial burden determined 3 days and 5 days post infection. The reduction in viability, expressed as log reduction, was determined by subtracting the bacterial burden in untreated mice and mice treated with tobramycin for 3 (C) or 5 days (D). Error bars indicate standard deviation. ****, significantly different (P<0.0001) from ΔsagS:: sagS.

Mutation of residue D105A enhances the susceptibility of P. aeruginosa to tobramycin treatment in the murine chronic pneumonia model.

Given the correlation between biofilm formation and bacterial load, we next asked whether SagS furthermore contributes to the persistence of biofilms when challenged with antimicrobial agents. We therefore exposed biofilms that were able to form during 24h of infection in the murine chronic pneumonia model to tobramycin. Considering that biofilms by ΔsagS::sagS and ΔsagS::sagS_L154A were previously found to be tolerant to tobramycin while biofilms by ΔsagS and ΔsagS::sagS_D105A were eradicated by 150 µg/µl tobramycin in vitro (Dingemans et al., 2018b), as determined using biofilm-MBC assays (Monzon et al., 2001, Villain-Guillot et al., 2007, Moriarty et al., 2007), we expected ΔsagS::sagS_D105A to have a disadvantage in maintaining a persistent infection compared to ΔsagS::sagS and ΔsagS::sagS_L154A.

Tobramycin treatment coincided with a low bacterial load (~1 log₁₀CFU) in the lungs infected with ΔsagS 3 days post infection while an average bacterial load of ~6.5 log₁₀CFU was noted for lungs infected with ΔsagS::sagS (Fig. S4A). Relative to untreated mice, tobramycin treatment had little effect on the bacterial load by ΔsagS::sagS, as tobramycin treatment only resulted in a reduction of the bacterial load by <1-log (Fig. 7C). However, tobramycin treatment had a more pronounced effect on the bacterial load by ΔsagS::sagS_D105A. Relative to no treatment (Fig. 7A), tobramycin treatment resulted in a significant 4-log reduction in the bacterial load by ΔsagS::sagS_D105A from 6.3 log₁₀CFU/lung to 2.1 log₁₀CFU/lung (Fig. 7C, S4A). While the overall bacterial load by ΔsagS::sagS_L154A was reduced relative to ΔsagS::sagS (Fig. 7A, S4A), tobramycin treatment had little effect on the bacterial load by ΔsagS::sagS_L154A, and only decreased from 4.2 log₁₀CFU/lung to 3.9 log₁₀CFU/lung, representing an overall reduction by 0.3-log (Fig. 7C, S4A). A similar trend was observed upon continued treatment 5 days post infection (Fig. 7D, S4B). Considering that biofilms by ΔsagS and ΔsagS::sagS_D105A were found to be impaired in drug tolerance in vitro, our findings suggest that persistence of P. aeruginosa biofilms in the murine lung to treatment by the antibiotic tobramycin correlates with in vitro biofilm drug tolerance phenotypes.

DISCUSSION

Due to their recalcitrance to antibiotics, P. aeruginosa biofilm infections are extremely difficult to treat, and have been associated with numerous medical conditions including periodontal disease, endocarditis, osteomyelitis, cystic fibrosis and indwelling medical devices (Costerton et al., 1999). In P. aeruginosa, the transition of the free-living to sessile lifestyle is orchestrated via the actions of different QS systems and two-component systems as well as factors and regulators contributing to an increase in the intracellular levels of cyclic-di-GMP (Fazli et al., 2014, Sakuragi & Kolter, 2007, Jimenez et al., 2012, Francis et al., 2017, Jenal et al., 2017, Winstanley et al., 2016). Many of these factors and regulatory circuits involved in enabling the biofilm lifestyle have been shown to contribute to P. aeruginosa causing persistent infections that are refractory to treatment (Mulcahy et al., 2014, Rybtke et al., 2011, Gellatly & Hancock, 2013, Mah & O’Toole, 2001, Winstanley et al., 2016). While the findings clearly demonstrated the biofilm mode of growth to be the culprit of the recalcitrant pathogenicity phenotype of P. aeruginosa, little attention has been paid to how drug tolerance is associated with biofilm formation, as both phenotypes tend to be linked. Thus, the contribution of biofilm formation or drug tolerance alone to infection and the pathogenicity phenotype of P. aeruginosa is unknown.

Amongst the regulatory systems contributing to the switch from the planktonic to the biofilm mode of growth in vitro by P. aeruginosa is the hybrid sensor kinase/response regulator SagS. This protein plays a key role by (i) controlling biofilm formation via a phosphorylation cascade that starts with the phosphorylation of BfiS and by (ii) contributing to the increased tolerance of biofilms to antimicrobial compounds (Fig. 1) (Petrova & Sauer, 2011, Gupta et al., 2013, Gupta et al., 2014, Petrova et al., 2017, Dingemans et al., 2018b). We previously demonstrated that distinct sets of amino acids in the periplasmic sensory domain of SagS compose the two major regulatory circuits that contribute to biofilm formation and antibiotic tolerance (Fig. 1). Specifically, our findings indicated that amino acid residue L154 in the HmsP sensory domain of SagS is involved in biofilm formation while amino acid residue D105 contributes to antibiotic tolerance. The findings provided the unique opportunity to address the question whether biofilm formation and/or biofilm drug tolerance contribute to the virulence phenotype by P. aeruginosa.

In this study, we therefore focused on SagS and two SagS variants harboring mutations in D105 and L154, that are key to the biofilm formation and antibiotic tolerance pathways, respectively, to elucidate the contribution of biofilm formation and biofilm drug tolerance to pathogenicity. Relative to the complemented strain ∆sagS::sagS, deletion of sagS which impaired both biofilm formation and biofilm drug tolerance, or alanine substitution of D105 and L154 had little to no effect on the expression of selected virulence genes and on P. aeruginosa virulence in acute infections as determined using a lettuce leaf model. Although the findings suggested that SagS contributes little to early signs of infection or lethality, SagS was found to contribute to P. aeruginosa fitness as determined using a murine acute pneumonia model. The fitness disadvantage became apparent during the later stages of acute infection in a murine model. Considering that SagS contributes to the transition to the surface-associated mode of growth in vitro (Petrova & Sauer, 2011), our findings suggest that in vivo, SagS likely assists in the acute-to-chronic virulence switch accompanying the motile-to-sessile mode of growth switch. Our findings of SagS playing a role in the acute-to-chronic switch is in agreement with the role of other regulatory factors and pathways. Moreover, given that only inactivation of sagS, but not alanine substitutions of L154 and D105, affected later stages of acute infection, our findings furthermore suggest that only one of the SagS-regulated phenotypes, biofilm formation or biofilm drug tolerance is sufficient to establish late-stage acute infections. In contrast, impairment of both biofilm formation and biofilm drug tolerance was found to coincide with reduced fitness in establishing late-stage acute infections. Overall, our findings suggest SagS to contribute to the acute-to-chronic switch in a manner similar to other regulatory factors and pathways previously reported to play a central role in the motile-to-sessile mode of growth switch and virulence. For instance, the RetS-LadS-GacSA-Rsm regulatory cascade has previously been reported to play a central role in biofilm formation in vitro and to contribute to the acute-to-chronic switch in vivo (Lapouge et al., 2008). Likewise, the transcriptional regulator AmpR has been shown to contribute to biofilm formation and biofilm drug tolerance by P. aeruginosa, with loss of ampR resulting in many phenotypes resembling a chronic infection strain (Balasubramanian et al., 2015).

Although our findings are in support of the idea that factors being involved in the motile-to-sessile switch also contribute to the acute-to-chronic switch in virulence, what sets SagS apart from other regulatory cascades is that the acute-to-chronic virulence switch likely only requires one of the biofilm-related phenotypes (Fig. 6), as in vivo persistence appears to depend on the ability of P. aeruginosa to establish biofilms. This is supported by the finding that mutation of residue L154, but not D105, lead to a significant reduction in bacterial load after 3 and 5 days of colonization in vivo (Fig. 7A-B). The findings of biofilm formation playing a role in vivo is in agreement with the role for SagS and the amino acid residue L154, but not D105, in mediating the transition from the planktonic to the biofilm mode of growth in vitro (Petrova & Sauer, 2011, Dingemans et al., 2018b). However, biofilm drug tolerance was furthermore found to contribute to in vivo persistence. This is supported by the noted decrease in the bacterial titer for ΔsagS::sagS_D105 5 days relative to 3 days post infection. It is likely that interference with the regulatory circuit leading to antibiotic tolerance, rendered P. aeruginosa more susceptible to antimicrobial compounds produced by the murine immune system such as neutrophil-derived ROS (Craig et al., 2009, Nguyen et al., 2017). Indeed, previously it was found that mutation of brlR renders P. aeruginosa biofilms susceptible to hydrogen peroxide (Liao & Sauer, 2012). Furthermore, Bayes et al. found that pulmonary neutrophilic inflammation was detected as soon as 48h post infection and persisted up to two weeks post infection in a murine chronic pneumonia model (Bayes et al., 2016).

Our findings furthermore demonstrate that recalcitrance to antibiotic treatment is primarily due to the biofilm drug tolerance phenotype, not the biofilm formation phenotype. This was supported by the finding that the most significant reduction in the bacterial load was detected for ΔsagS::sagS_D105 compared to ΔsagS::sagS and the ΔsagS::sagS_L154A variant (Fig. 7C-D). The findings suggested that recalcitrance to antibiotic treatment is primarily due to biofilm drug tolerance. It is of interest to note that while mutation of residue L154, but not D105, lead to a reduction in bacterial load in vivo using a chronic pneumonia model, with mutation of residue D105 rendering in vivo biofilms significantly more susceptible to tobramycin, very little to no difference in virulence gene expression was noted amongst the ΔsagS mutant strains. This finding was surprising considering that the transition to chronic infections has been reported to coincide with a switch from acute virulence factors (swarming motility, lipase, rhamnolipids, and type III secretion) to chronic virulence factors with the latter including the ability to form biofilms but also type VI secretion, and QS-regulated factors more traditionally thought of as involved in acute infections, such as pyocyanin and hydrogen cyanide (Winstanley et al., 2016). While we do not exclude the possibility of SagS contributing to a distinct set of virulence genes not tested here, our findings instead support the notion that the ability to form biofilms and/or P. aeruginosa biofilms being rendered resistant to antimicrobial agents are the main contributor to the P. aeruginosa virulence phenotype.

Taken together, our findings indicate distinct roles for biofilm formation and biofilm drug tolerance in the virulence phenotype of P. aeruginosa. Moreover, given the effect of alanine substitution of L154 on bacterial load and persistence in the murine lung, our findings suggest that chronic infections more closely resemble a biofilm growth mode. In addition, our findings suggest that in vitro biofilm drug tolerance resembles the refractory nature of chronic infections, apparent by the in vivo recalcitrance to antimicrobial agents, and may be the reason why conventional therapies have proven inadequate in the treatment of many (if not most) chronic biofilm-related infections. Moreover, our findings indicate that late stage acute infections and the likely acute-to-chronic virulence switch relies on the ability of P. aeruginosa to form biofilms or transition to a drug tolerance state, with impairing both abilities resulting in the acute-to-chronic virulence switch being disabled. The finding of distinct sets of amino acids in the periplasmic sensory domain of SagS such as L154 and D105 contributing to biofilm formation and antibiotic tolerance both in vitro and in vivo furthermore suggests SagS to be a suitable target for a potential anti-biofilm treatment strategy. Such a novel anti-biofilm treatment strategy may be based on strategically interfering with SagS signaling by blocking the key residues D105 and L154, mimicking alanine substitutions, which will effectively impair the two divergent regulatory circuits of SagS.

EXPERIMENTAL PROCEDURES

Bacterial strains, plasmids, and culture conditions

All bacterial strains and plasmids are listed in Table 1. Overnight cultures were grown in Lennox Broth (LB) at 37°C under shaking conditions (220 rpm). Antibiotics for plasmid selection or maintenance were used at the following concentrations: 60 μg/mL tetracycline and 250 μg/mL carbenicillin for P. aeruginosa; and 20 μg/mL tetracycline, 50 μg/mL kanamycin and 100 μg/mL ampicillin for Escherichia coli (E. coli).

Table 1.

List of strains and plasmids used in this study.

Strains/plasmids	Relevant genotype or description	Source
Strains
Escherichia coli
NEB 5-alpha Competent E. coli (High Efficiency)	fhuA2 (argF-lacZ)U169 phoA glnV44 80 (lacZ)M15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17	New England Biolabs
SM10	thi-1 thr leu tonA lacY supE recA::RP4–2-Tc::Mu; KmR	(de Lorenzo & Timmis, 1994)
Pseudomonas aeruginosa
PAO1	Wild type	(Holloway, 1955)
ΔsagS	PAO1, ΔsagS (PA2824)	(Petrova & Sauer, 2011)
ΔsagS::CTX	ΔsagS harboring the empty pMini CTX vector, TetR	This study
ΔsagS::sagS	ΔsagS harboring chromosol insertion of sagS under the control of the PsagS promoter at attB site, cured pMini CTX vector	This study
ΔsagS::sagS_D105A	ΔsagS harboring chromosol insertion of sagS_D105A under the control of the PsagS promoter at attB site, cured pMini CTX vector	This study
ΔsagS::sagS_L154A	ΔsagS harboring chromosol insertion of sagS_L154A under the control of the PsagS promoter at attB site, cured pMini CTX vector	This study
Plasmids
pMJT-sagS	C-terminal HA-tagged wild-type sagS cloned into pMJT1 at NheI/SacI; AmpR (CarbR)	(Petrova et al., 2017)
pMJT-sagS_D105A	C-terminal HA-tagged sagS harboring a D105A substitution in pMJT1 at NheI/SacI; AmpR (CarbR)	(Dingemans et al., 2018b)
pMJT-sagS_L154A	C-terminal HA-tagged sagS harboring a L154A substitution in pMJT1 at NheI/SacI; AmpR (CarbR)	(Dingemans et al., 2018b)
pMini-CTX-lacZ	attP site-specific integration vector; TetR	(Becher & Schweizer, 2000)
PsagS-sagS	pMini-CTX harboring wild-type sagS under control of the sagS promoter region (1–460 bp upstream of the sagS start codon)	This study
PsagS-sagS_D105A	pMini-CTX harboring sagSD105A under control of the sagS promoter region (1–460 bp upstream of the sagS start codon)	This study
PsagS-sagS_DL154A	pMini-CTX harboring sagSL154A under control of the sagS promoter region (1–460 bp upstream of the sagS start codon)	This study
pRK2013	Helper plasmid for triparental mating; mob; tra; KmR	(Figurski & Helinski, 1979)
pFLP2	Plasmid used for FLP-mediated recombination; ApR (CarbR)	(Hoang et al., 1998)

Strain construction

The sagS gene as well as its two site-directed mutant variants sagS_D105A and sagS_L145A, were cloned under the control of the P_sagS promoter (promoter region comprised of 1–460 bp upstream of the sagS start codon) into the pMini-CTX-lacZ vector using the DNABuilder® HiFi DNA Assembly cloning kit (New England Biolabs). Briefly, the sagS gene or its mutant variants (sagS_D105A and sagS_L145A) and P_sagS were PCR-amplified using the Q5 Hot-Start High-Fidelity 2X Master Mix (New England Biolabs) following the manufacturer’s guidelines. Moreover, pMini-CTX-lacZ was digested using the EcoRI-HF and HindIII-HF restriction enzymes (New England Biolabs). Then, the assembly reaction was performed for 15 min at 50°C using a vector to insert ratio of 1:2. This was accomplished by combining 0.013 pmol of digested vector with 0.026 pmol of both the sagS promoter PCR product and the sagS PCR products. The assembled constructs (2 μl) were subsequently transformed into E. coli DH5α (New England Biolabs). The identity of vector inserts was confirmed by PCR using sagS_NheI/sagS_HA_SacI rev primers (Table 2) and sequencing using CTX-sagS_promoter_F/ sagS_NheI_for primers (Table 2). The resulting plasmids were introduced into ΔsagS via triparental mating using E. coli harboring pRK2013 as the mobilizer strain, after which proper integration of the pMini-CTX vector was verified via PCR using Pser-up/Pser-down primers (Table 2). To generate unmarked complemented ΔsagS mutant strains, the pMini-CTX backbone including the tetracycline resistance and integrase genes was first removed by FLP recombinase-mediated excision (Hoang et al., 1998), and the pFLP2 plasmid cured via sucrose counterselection.

Table 2.

Oligonucleotides used in this study.

Oligonucleotide	Sequence
Cloning
CTX-sagS promoter_F	AGGATCCCCCGGGCTGCAGGAATTCGGCCAGGTGGTCGCGCTC
CTX-sagS promoter_R	CGCCTAGCATAGCCCATCCCGACCGTCTG
CTX-sagS_F	GGGATGGGCTATGCTAGGCGGCAGAACC
CTX-sagS_R	CGAGGTCGACGGTATCGATAAGCTTCTAGTCGCTCGCGGTGAG
PCR-Screening and Sequencing
Pser-up	CGAGTGGTTTAAGGCAACGGTCTTGA
Pser-down	AGTTCGGCCTGGTGGAGCAACTCG
sagS_NheI for	GCGCGCGCGCTAGCATGCTAGGCGGCAGAACCTCGC
sagS_HA_SacI rev	GCGCGCGCGAGCTCCTAAGCGTAGTCTGGGACGTCGTATGGGTAGTCGCTCGCGGTGAGCGG
RT-qPCR
mreB_F	CTTCATCAACAAGGTCCACGA
mreB_R	GCTCTTCGATCAGGAACACC
cysD_F	CTGGACATCTGGCAATACAT
cysD_R	TCTCTTCGTCAGAGAGATGC
chiC_F	CAGTTGCACCAGGCCCGC
chiC_R	GGCCAGGTCGATGGCGC
vqsR_F	CTCCGAAGATTTCGAAACGA
vqsR_R	GGCATAGGGTTCTTTCACCA
amrZ_F	GAGCAGATCGCAGAAGTCG
amrZ_R	GCGAACACCGAGATTGTCTT
lasB_F	GCCGCCGACCTGATCG
lasB_R	CAGCACCTGCTCGGCG
rhlA_F	GCGCGATGGCGACCAC
rhlA_R	CACCACCGAGCTGCGG
toxA_F	CCTCAGCATCACCAGCGAC
toxA_R	GCAGGCGATGACTGATGAC
hcnA_F	TGAACGTCAACACGATATCCA
hcnA_R	CATTGAGCACGTTGAGCAC
phzB_F	TTCCTGCATTCCTTCGAACT
phzB_R	ATGCCTTCGCGTTTGATCT
pcrV_F	GACCCCACGCTATATGGCTA
pcrV_R	GAGCCGCTGAGAAAATCCTT
pscL_F	CGCGACTACCAGGACTACCT
pscL_R	CGCTTCTGCTCCTGGTAAAC
exsA_F	ATCGAGGAGTTGCTGATGCT
exsA_R	TCCATGAATAGCTGCAGACG
exoS_F	TGGTCTCTACACCGGCATTC
exoS_R	CCTTGGTCGATCAGCTTTTG
fliC_F	TGCAGCAGTCCACCAATATC
fliC_R	TCGGAGATACGGGTCAGTTC
pilQ_F	ACCTGGAGAAACTCGACGTG
pilQ_R	CGGCTGCTCGATGGTATAG
glpD_F	GAAACAGAAGTCGCCCTA
glpD_R	GGTGAAACGGTCCAGATAGG
betI_F	GAAGGCAACTTCGACGACA
betI_R	GAATACAGGCGATGGTCGTT
norC_F	CCGAGACCTTTACCAAAG
norC_R	CAGGGTCTTCTCGGTG
dnr_F	GTTCCGTTTCTCCAACAAGG
dnr_R	GAAAGCGTCTCGATCTCGTC
anr_F	GCAACGAGATCGGCAACTAT
anr_R	CTCGATGGAGTCGAGGATGT
pilY1_F	CAATTGCTGAACGACTCGAA
pilY1_R	AAGAAGTTCACCCGGTGTTG
pelA_F	GGTGCTGGAGGACTTCATC
pelA_R	GGATGGCTGAAGGTATGGC
pslG_F	CACGTAAGGGACTCTATCTGG
pslG_R	AGGAAGTCTTTCCAGACCAC

Growth curves

Overnight cultures of P. aeruginosa grown in LB medium were washed twice with saline before being inoculated in different media (swarming, swimming, phosphatidylcholine, and LB medium) at a 1:100 dilution. The swarming medium consisted of M8 minimal medium (Na₂HPO₄ 6 g/L, KH₂PO₄ 3 g/L, and NaCl 0.5 g/L) supplemented with 0.5% casamino acids, 0.2% glucose and 1mM Magnesium sulfate. Swimming medium contained 0.1% tryptone and 0.5% sodium chloride. Phosphatidyl choline medium consisted of M9 minimal medium (Na₂HPO₄ 12.8 g/L, KH₂PO₄ 3 g/L, NaCl 0.5 g/L, and NH₄Cl 1 g/L) supplemented with 0.05% (w/v) phosphatidylcholine and 1% (w/v) Brij58 (Spectrum Chemical MFG Group). Growth curves were determined by measuring the absorbance at 595nm every 30 min with 10 s shaking before each measurement for a total time period of 24h (swarming, swimming, and LB medium) or 48h (phosphatidylcholine medium) using a SpectraMax i3x plate reader (Molecular Devices).

Biofilm formation

To determine the biofilm architecture, biofilms were grown in 24-well plates in 5-fold diluted LB medium under shaking conditions (220 rpm), with the growth medium being exchanged every 12h. Confocal laser scanning microscopy (CLSM) images were acquired using a Leica TCS SP5 confocal microscope (Leica Microsystems, Wetzlar, Germany). Prior to microscopy, biofilms were stained using the LIVE/DEAD BacLight Bacterial Viability Kit (Life Technologies). Quantitative analysis of the confocal laser scanning microscope images of 24-well plate-grown biofilms was performed using COMSTAT analysis (Heydorn et al., 2000). For RNA isolation, biofilms were grown at 22°C in 20-fold-diluted LB medium, using a continuous flow tube reactor system (1-m-long size 13 silicone tubing; Masterflex, Cole Parmer, Inc.) with an inner surface area of 25 cm² at a flow rate of 0.1 ml/min, as previously described (Sauer et al., 2002, Sauer et al., 2004, Petrova & Sauer, 2009).

RNA isolation and qRT-PCR

For RNA isolation, biofilms grown in biofilm tube reactors were harvested by extrusion, with the cell paste being collected directly into 500 μl of RNAProtect Bacteria Reagent (Qiagen). Planktonic cells were grown to mid-exponential growth phase, and subsequently mixed with RNAProtect Bacteria Reagent at a 1:1 ratio. Thus obtained samples were stored at −80°C until RNA extraction. Total RNA was purified using the RNeasy Mini Kit (Qiagen), with residual DNA being removed using 1 µl of (2U/µl) Turbo DNase (Ambion) for 30 min. cDNA was prepared using the iScript Select cDNA synthesis kit (Biorad) using 1 µg RNA. qRT-PCR was performed using the BioRad CFX Connect Real-Time PCR Detection System and SsoAdvanced SYBR Green Supermix (BioRad) with primers listed in Table 2. The cysD and mreB genes were used as housekeeping genes under planktonic and biofilm growth conditions, respectively. Fold changes were calculated using the Livak method (Livak & Schmittgen, 2001). Melting curve analyses were performed to verify specific single product amplification.

Motility assays.

Swimming assays were performed using medium composed of 0.1% tryptone, 0.5% sodium chloride and 0.3% agar as previously described (Toutain et al., 2005). Swarming assays were performed using M8 minimal medium (Na₂HPO₄ 6 g/L, KH₂PO₄ 3 g/L, and NaCl 0.5 g/L) supplemented with 0.5% casamino acids, 0.2% glucose, 1mM Magnesium sulfate, and 0.4% agar as previously described (Southey-Pillig et al., 2005, Merritt et al., 2010). All media were inoculated by stab-inoculation. For both assays, the diameter of growth from the point of inoculation was measured using a ruler.

Virulence testing using plant infection models

The role of SagS and its mutant variants in virulence was assessed using two plant infection models. The lettuce model allows for the analysis of acute virulence. Lettuce leaf infection was carried out essentially as described by Filiatrault et al. (Filiatrault et al., 2006) using store-bought romaine lettuce leaves, which were each infected by injection of 100 µl of an overnight culture of P. aeruginosa that was diluted to an OD_600nm of 0.2, washed twice, and resuspended in 10 mM MgSO₄. Leaves, placed in a plastic bags containing a water-saturated tissue paper, were subsequently incubated at 37°C, and the spread of infection along the midrib recorded daily. The A. thaliana infection model provides a quantitative approach and permits the tracking of bacterial cell proliferation in planta (Starkey & Rahme, 2009). A. thaliana were grown from seeds. Sterilized seeds were first transferred onto ½ MS medium (2.2 g/L Murashige and Skoog basal medium, 1.5% sucrose, 0.5g/L MES) containing 0.1% agar, and incubated for 2 days at 4°C in the dark to synchronize seed germination. Then, seeds were incubated for a period of up to 12 days at a 25/22°C, with a 16h-light/8h-dark cycle. After 8 or 12 days, plants were transferred to 24-well plates containing 1.5 ml of ½ MS medium, and incubated for two days at 25/22°C with a 16h-light/8h-dark cycle prior to infection by P. aeruginosa mutant strains having a final optical density at 600 nm (OD600) of 1.0. This was accomplished by adding 100 µl of P. aeruginosa concentrated to an optical density of 16 at 600 nm to each well. Plants were subsequently incubated for 14 days in a 25/22°C, 16h-light/8h-dark cycle and inspected daily for infection (first sign of discoloration) and plant death (discoloration of the entire plant).

Ethics statement

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Illinois at Urbana-Champaign (Protocol Number: 15171).

Murine model of acute pneumonia infection

Acute pneumonia infection was performed as previously described (Lau et al., 2004). Six-week old male and female adult CD-1 mice (in cohorts of 10) were intranasally-infected with 10⁷ CFU of each P. aeruginosa strain. Infected lungs were homogenized for bacterial burden determination at designated times. Attenuation is defined as the log₁₀ difference in CFU between wild-type and mutant bacteria recovered from lung tissue 16 h after inoculation.

Murine model of chronic pneumonia infection

Chronic mouse infections were performed as previously described (Li et al., 2014). Briefly, stationary phase P. aeruginosa bacteria were entrapped in agarose beads by mixing with heavy mineral oil at 52°C (Sigma, St. Louis, MO) and stirred vigorously for 6 min followed by cooling on ice for 10 min. The bacteria-containing beads were centrifuged (9,000 g for 20 min at 4°C) followed by extensive washing in PBS. The beads were passively filtered through sterile 200-µm-diameter hole nylon mesh and verified for size (70–150 µm diameter) and uniformity by microscope examination. An aliquot of beads was homogenized and plated onto LB agar plates to determine the bacterial CFU. A 50 µl inoculum containing 10⁶ CFU of P. aeruginosa was introduced into the lungs of adult BALB/c mice (7-week old, cohorts of 10) via the trachea non-surgically by using a 21-gauge ball-ended needle. Successful delivery of the beads to the lungs was manifested by choking reaction of the challenged mice immediately after instillation followed by rapid breathing. Mouse lungs were homogenized for bacterial burden at designated times.

Statistical analysis

To determine statistical differences between strains, a one-way ANOVA was performed followed by a Dunnett’s post-hoc test using Prism5 software (Graph Pad, La Jolla, CA, USA).